AI is increasing the cost of every minute of compromise, says Coates cyber & risk leader

Strong perimeter controls cannot protect operations once an attacker gains access.

As AI increases the speed and scale of cyber attacks, organisations have less time to detect suspicious activity, contain exposure and recover safely.

In an interview at Security Edge, Mohamed Omran, Head of Cyber Security and Risk at Coates, joins ADAPT’s Head of Programs & Value Engagement Byron Connolly to discuss resilience across a distributed, digitally connected business.

He shares how Coates links detection, response and recovery with practical governance across suppliers, operational technology and AI agents.

Key takeaways:

• Cyber resilience requires rapid detection, containment, recovery and post-incident learning alongside strong preventative controls.
• AI allows attackers to run more actions concurrently, increasing the volume of activity security teams must analyse and contain.
• Secure innovation requires clear supplier expectations, named risk ownership and human oversight when AI agents can act or access sensitive data.

Resilience extends beyond the perimeter

Effective cyber defence assumes that some attacks will get through.

Authentication, email gateways and other preventative controls remain important.

Once an attacker crosses the perimeter, detection, rapid lockout and limited access determine how far the incident can spread.

Coates plans for the full attack lifecycle, including response, recovery and post-incident learning.

This reduces the operational and data impact of phishing, business email compromise and other successful attacks.

Recovery planning also gives the organisation a tested route back to safe operations rather than leaving critical decisions until an incident is underway.

Defence must match the speed of AI attacks

AI allows attackers to run more actions at once and move across infrastructure or network perimeters faster than traditional security processes can respond.

Even a short compromise can expose significantly more data.

Security teams face a larger volume of activity to analyse, prioritise and contain within a compressed window.

Mohamed sees a role for AI-supported analysis and response, combined with human judgement.

Technology can identify patterns and contain threats at machine speed, while people assess context, consequences and the appropriate response.

Capabilities from security partners can help close that speed gap when they are evaluated through carefully tested private previews before wider deployment.

Shared ownership supports secure innovation

Coates operates distributed sites with tens of thousands of connected devices.

Its security risks extend across office IT, operational technology, suppliers, business teams and AI agents.

Clear contractual expectations, secure identities, patching and threat modelling help protect data and physical safety across that ecosystem.

Named owners establish accountability when risks span several teams or external providers.

Mohamed describes the approach as “yes, securely.”

Teams can test new AI tools in a sandbox, scan code and packages through a repository and define ownership before moving towards production.

Human escalation remains essential when an agent can act or access sensitive data.

These controls give teams room to innovate within boundaries tied to the business use case and its risk.